LF-3270 Test Plan

https://lite-farm.atlassian.net/browse/LF-3270

Currently JWT expiration is set at 7d.

Desired implementation:

Best practices to keep JWT tokens very short lived (e.g.: ~10 minutes) and to have them periodically refreshed using a single-use HTTP-only cookie (AKA: refresh token).

Area / scope to test	Front end	Back end	Model	Notes

Area / scope to test	Front end	Back end	Notes
Requirement specific constraints	Testing as per Jira ticket	To test this the user needs to follow several steps: Get a new access and refresh token Null out the refresh token. Hit an endpoint both before and after the access token expires. Ensure it works before and fails after.
Role based constraints			Does role determine what a user can see or do? Is this enforced uniformly across the front end and back end?
User preferences constraints			Is this impacted by user or farm preferences such as language, system of measure, certification status?
Numerical input constraints			Do we appropriately handle negative, very small, very large, or 0 as inputs?
Text input constraints			Do we appropriately handle blank, very small, and very large inputs? Is there a strict format (such as email) that must be followed?
Date based constraints			Are there logical restrictions on what dates can be input? Should a use be able to complete something in the future for example.
Date based assumptions			Are we making valid assumptions about what dates should be allowed?
Timezone driven interactions			If timezones play a role in the data, are they being displayed or converted appropriately?
Interaction / transitioning UI based constraints			Is the UI transitioning appropriately? Is the API providing da
Flow based constraints			Is state being preserved appropriately in a flow? If I go back and then forth, is it maintained? Is state invalidated when it should be?
Synchronous / asynchronous constraints			Is the interaction synchronous, asynchronous, or does it support both? Can you simulate both if so?
Time-out / low bandwidth constraints			Does the feature fail gracefully under no bandwidth / low bandwidth environments?
Data transformation correctness			Are values appropriately updated when units change? Is it WYSIWYG?
Outcome correctness			When inputting known inputs with expected outputs - do you get the results you expect? Have you tested several “cases” of this?
Switching farms			Does this feature respond well to switching farms (and returning)?
Notification constraints			Should a notification be marshalled based on this action?
Cascading effects			Are there logical places
Integration constrains			Do we need to ensure state is consistent between LiteFarm and the external service? What failure cases do we need to handle? How do we report back the outcome to the user or external service?
Concurrency			How do changes made to records affect other users on the farm? e.g. What happens when a record is soft deleted while another user is viewing said record?